A model towards global demographics: an application—a universal bank branch geolocator based on branch size

Abstract

Branch size strongly depends on branch cash holdings. However, while any exhaustive study into branch cash holdings1 must include demographics around branches, there are major variations when defining demographics according to “local” parameters, as opposed to “internationally accepted” ones. This wide fluctuation in definitions makes cross-border comparisons very difficult. The present paper intends to overcome these difficulties by developing a global spatial model that uses cash holdings as a major determinant of branch size and where geographical concepts are replaced by “internationally accepted” notions. Specifically, the contributions of this paper are twofold: firstly, it presents a theoretical model (based on Markov and Gibbs random fields) to analyse the branch cash holdings from a global spatial standpoint. Secondly, it introduces a universal branch geolocator based around a decision model that redesigns the bank branch network according to branch size. Importantly, the model variables (including branch size as the main criterion) can be replaced/expanded as needed through the use of a highly versatile decision-making tool that can be applied to a wide range of contexts, even non-banking ones as long as they are influenced by demographics.